Regularization of Predictors of GDP and Individual Sectors of the Economy of the Republic of Kazakhstan

Abstract

The purpose of the study is to identify new, alternative directions for the development of GDP and individual industries in the Republic of Kazakhstan.

Methodology - Applying a limited number of predictors and using obvious patterns, it is possible to build statistically significant models of economic development in a country or region. In the case of Kazakhstan, such models have shown their practical failure. This study proposes an alternative approach aimed at covering as many predictors as possible and abandoning a priori theories and judgments.

In practical terms, the compression (regularization) of predictor set was carried out using the LASSO method (least absolute shrinkage and selection operator). The statistical significance of the selected predictors was investigated using the least squares method. The models were improved using the backward elimination method.

Models of development of GDP, stock market and civil aviation of the Republic of Kazakhstan have been built. The research data frame consists of 8 blocks: socio-demographic indicators, living standards, labor market and wages, prices, national economy, real sector of the economy, trade, financial system, as well as data on the capitalization of the Kazakhstan Stock Exchange.

Novelty (value) of the research - models have been developed for working with high-dimensional data, which are features of developing countries.

Results of the study - according to the results of the study, the stock market index turned out to be sensitive to a wide range of social and macroeconomic indicators: population growth, unemployment, poverty, inflation, investment, devaluation. Our conclusion: the development of the stock market does not require any specific financial measures, it is necessary to deal with the economy as a whole. The volumes of transactions in corporate securities do not have stable external predictors. The main indicators of the republic's civil aviation have stable external predictors. Passenger turnover, passenger dispatch directly depend on: unemployment rate, wages, GDP per capita, various types of services and products, money supply. There are no external predictors for a separate type of aviation work - cargo transportation. Consequently, positive results can be obtained by reforming this particular segment of services.